Evolutionary Algorithms Application in Portfolio Problem Selection

Abstract

This paper focuses on portfolio selection problem solving through the application of some evolutionary algorithms such as genetic algorithms (GA), particle swarm (PS) and differential evolution (DE).Therefore we focus mainly on the performance of genetic algorithms compared to the other two techniques on data collected from the Tunis Stock Exchange over a period of 5 years. We find in this context that genetic algorithms are similarly more efficient in choice of the optimal of portfolio compared to PS and DE.